AlexNet
This project intends to replicate the AlexNet CNN network by strictly following the original paper.
1. Dataset
Currently, the project is configured to be trained on imagenette/full-size-v2, a dataset delivered by Tensorflow. From the documentation:
Imagenette is a subset of 10 easily classified classes from the Imagenet dataset. It was originally prepared by Jeremy Howard of FastAI. The objective behind putting together a small version of the Imagenet dataset was mainly because running new ideas/algorithms/experiments on the whole Imagenet take a lot of time.
Link: https://github.com/fastai/imagenette
TODO: Find ImageNet 2010
Please drop me an email if you know where to find this specific dataset
2. Prerequisites
- Python 3.7.9
- Tensorflow 2.3.1
pip install tensorflow-gpu==2.3.1
3. Setup
- Clone the repo:
git clone https://github.com/houseofai/alexnet.git
- Install the 3rd party packages
cd alexnet/
pip install -r requirements.txt
4. Training
To train the model, launch the following command:
python train.py
To test the configuration on a smaller dataset:
python train.py --conf=test
5. Predict
To predict class probabilities on an image:
python train.py --mode=predict --image=/path/to/image